CLONAL VARIATION OF WINTER COLD HARDINESS AMONGST VITIS VINIFERA A Report Presented to the Faculty of the Graduate School Of Cornell University In Partial Fulfillment of the Requirements for the Degree of Master of Professional Studies by Michael Quade August 2021 © 2021 Michael Quade ABSTRACT The cold hardiness and deacclimation rates of various clones of four V. vinifera grapevine varieties were evaluated by measuring lethal temperatures of dormant buds using low temperature exotherms. This study aimed to elucidate the potential cold hardiness trait differences across clones of each variety. Levels of cold hardiness play an important role in grower choice when considering planting new fields or replacing lost crops in regions subject to low temperatures during the winter. Ultimately, it was shown that there is no significant cold hardiness difference between clones within any of the four sampled varieties. However, a revised experimental design and additional sampling must be done over multiple seasons to gain a better view of any potential clonal variation that may arise. It is likely that the microclimatic environmental differences between vineyard sites was a compounding factor, and as such the effect of clonal variation was challenging to isolate. The results of this study did show a discernable difference in cold hardiness levels and deacclimation rates between the sampled varieties, which is well-documented in other studies in this area. While it is still unknown whether clone has an influence on winter cold hardiness, these results may be important for future vineyard plantings in regions with increased winter temperature fluctuation as a result of ongoing climate change. Additionally, further investigation is warranted into the potential genetic variation between clones within a variety, as this may uncover the basis for other clonal trait variations of interest. BIOGRAPHICAL SKETCH Michael Quade was born in the Finger Lakes region of Upstate New York. He has had a life-long affinity for nature and the effect of the environment on trait expression. He graduated from the University of Rochester with a Bachelor of Science in Ecology & Evolutionary Biology in 2016. During his undergraduate studies, he conducted research with Dr. Brisson’s laboratory studying phenotypic responses to the environment in Acyrthosiphon pisum, the pea aphid. After graduation, he returned to work in the nursery at Hermann J. Wiemer Vineyard, and inspired by the burgeoning local grape industry, decided to pursue his master’s degree in Horticulture from Cornell University. It is here that he developed his capstone research project investigating the potential variations in cold hardiness amongst clones of Vitis vinifera grapevines. This topic particularly inspired him, as he has had countless correspondence with grape growers regarding the suitability of specific clones in their growing regions. The results of this study lie herein. ACKNOWLEDGMENTS I would like to express my gratitude to Dr. Bruce Reisch and Dr. Jason Londo for providing me the opportunity to work on this project and guiding me through concepts of experimental design, as well as the challenges of data analysis. I would also like to thank Mike Colizzi, Hanna Martens, and Lex Pike for their contributions during sample collection. I appreciate their patient explanations regarding methods. I will also be eternally grateful to all the professors and classmates I’ve met in the Horticulture program who helped me thrive during my time at Cornell University. iii TABLE OF CONTENTS ABSTRACT ....................................................................................................................................................... iii BIOGRAPHICAL SKETCH ............................................................................................................................ iii ACKNOWLEDGMENTS ................................................................................................................................. iii TABLE OF CONTENTS .................................................................................................................................. iv LIST OF ABBREVIATIONS ............................................................................................................................ v 1. Introduction ................................................................................................................................................ 1 2. Materials and Methods .............................................................................................................................. 4 2.1 Bud collection & Preparation ............................................................................................................... 4 2.2 DTA Freezer Runs ............................................................................................................................... 4 2.3 Peak Calling ......................................................................................................................................... 5 2.4 iButton Temperature Logging .............................................................................................................. 5 2.5 Statistical analysis ................................................................................................................................ 6 3. Results .......................................................................................................................................................... 6 3.1 LTE Measurements .............................................................................................................................. 6 3.2 DeAcclimation Rates ........................................................................................................................... 9 3.3 Temperature Logging Data ................................................................................................................ 11 4. Discussion .................................................................................................................................................. 12 4.1 Limitations ......................................................................................................................................... 12 4.2 Future Research .................................................................................................................................. 13 REFERENCES ................................................................................................................................................. 15 iv LIST OF ABBREVIATIONS CF, Cabernet Franc; CH, Chardonnay; PN, Pinot Noir; RS, Riesling; DTA, Differential Thermal Analysis; LTE, Low Temperature Exotherms; DeAcc, Deacclimation rate HJW, Hermann Wiemer Vineyard original plot; Magd, Magdalena Vineyard; Josef, Josef Vineyard; JV, Julia Vineyard; KV, Kasper Vineyard; v 1. Introduction Clonal choice of a vinifera grapevine variety is an important consideration and talking point amongst grape growers worldwide. These different clones have developed over centuries of selection, in which generational growers would choose their best-quality vines and propagate them via cuttings, eventually replacing entire fields with their designated “clone”. Over time, this clonal material gets further propagated and diffused into the market, with just a handful dominating the industry due to their widely desirable traits, such as cluster size, flavor, and yield. The pride and lore surrounding any clone runs so deeply that their classification has become a topic of hot debate in recent years. International organizations such as ENTAV-INRA and Foundation Plant Services have stepped in to develop labeling standards so that clear definitions of the clones of vinifera grapevines can exist, and the minutiae of trait differences between them can be catalogued for public knowledge. Even these organizations cannot always agree on the classification of a clone, and as such the industry divide and confusion only continues to grow. Using scientific tools, it becomes possible to discern the variations between clones, or lack thereof, more precisely than just with eyesight alone. Measuring complex phenotypes such as cold hardiness is nearly impossible for an average grower, and with the looming threat of climate change it becomes imperative that accurate information is publicly available regarding the future suitability of a specific clone in the region the grower is considering planting in. As the industry deals with conflicting sources, jumbled clone numbers, and unreliable word-of-mouth, it is left up to scientists to provide a trusted source of information to current and potential growers that will help guide their planting decisions for years to come. 1 In this study, a close look at the cold hardiness of various clones of four vinifera varieties was undertaken, (Table 1). These varieties were chosen due to their widespread contemporary use in the Finger Lakes Region, which allowed for multiple samplings across five vineyards along the Western side of Seneca Lake (Figure 1). ENTAV Origin Synonyms Sampling Location Variety Clone Maine-et-Loire, FPS11, imported in 1995 Julia, Magd 214 France 312 France FPS13, donated in 1998 Julia, Magd Gironde, France FPS12, import date Julia, Magd Cabernet 327 unknown Franc Pyrénées- FPS4, imported in 1988 Julia, Magd Atlantiques, 332 France Montpellier, FPS1, imported in 1938 Julia, Magd 3880 France Champagne FPS37, imported in 1988 HJW, Julia, Magd Perrier-Jouet, Chardonnay 95 France 96 Dijon, France FPS70, imported in 1988 Magd Champagne FPS73, imported in 1988 Kasper Perrier-Jouet, 115 France Pinot Noir Loire Valley, FPS20 Magd 348 France 667 Dijon, France FPS72, imported in 1992 Magd 777 Dijon, France FPS71, imported in 1992 Kasper Neustadt clone, FPS12, imported in 1963 HJW, Julia, Magd Rheingau, 90 Germany Geisenheim, FPS24, imported in 1952 Josef, Julia 110 Germany Geisenheim, Imported in 1950’s HJW, Julia Riesling 198 Germany Geisenheim, FPS23, imported in 1987 HJW, Julia, Magd 239 Germany Neustadt, FPS21, 356 Trautwein Josef, Julia 356 Germany “German” Unknown Josef Table 1. Clones of four vinifera varieties sampled at various vineyard sites 2 Figure 1. Map of Finger Lakes Region showing the vineyard sampling sites along Seneca Lake. To measure cold hardiness in these selected varieties/clones, dormant bud samples were collected across the 2020-21 winter season and using differential thermal analysis (DTA), the precise temperature of bud lethality was recorded. DTA is a method which uses an incremental reduction in temperature via a programmable freezer to apply freezing stress to the buds until they are overwhelmed and die off. This exact moment is measured via thermoelectric modules which record the bud death as a momentary “peak” of electrical conductivity due to a rapid expulsion and freezing of water from the dying bud (Mills et al., 2006). Using the resulting set of temperature peaks, it becomes possible to extrapolate the cold hardiness of a specific clone, as well as the rate at which they will deacclimate from dormancy at a given time throughout the season (Alvarez et al., 2018). This deacclimation rate is equally important for grower decision- making in terms of clonal selection, as a clone or variety that deacclimates quickly may experience bud break early in the spring, when there is still a high risk of frost damage. 3 2. Materials and Methods 2.1 Bud collection & Preparation Dormant cane samples were collected during four timepoints throughout the winter of 2020-21. The timepoints of collection and DTA freezer runs can be found in Table 2. Each cane held 10 dormant leaf and shoot buds, and five canes of each clone being tested were collected, for a total of 50 buds per clone at each timepoint. After collection, the canes were brought back to the AgriTech laboratory and cut into individual lengths of approximately 10cm, each containing an individual or double bud. Following this, the buds were separated into three categories: • Initial LTE – 10x buds to be run in the freezer immediately to establish a baseline of LTE at collection • De-Acclimation Rate – 25-30x buds which are left at 25oC and run in groups of 5x buds over consecutive days to plot the curve of dormancy reduction over time • Bud Burst – 10x buds to be left at room temperature until bud dormancy is broken, to compare the observed vs. expected rate values Initial Run 2nd Run 3rd Run 4th Run 5th Run 6th Run Collection 1 (C1) 12/1/2020 12/4/2020 12/8/2020 12/13/2020 12/16/2020 Collection 2 (C2) 1/14/2021 1/16/2021 1/19/2021 1/22/2021 1/26/2021 1/29/2021 Collection 3 (C3) 2/25/2021 2/27/2021 3/2/2021 3/5/2021 Collection 4 (C4) 3/26/2021 3/27/2021 3/28/2021 3/29/2021 3/30/2021 3/31/2021 Table 2. Dates of bud collections and subsequent DTA freezer runs. 2.2 DTA Freezer Runs To prepare for a freezer run, each bud was dissected away from the cane length, and kept in separate piles depending on their clone designation. Specialized trays with 9 separate thermoelectric modules (“wells”) were prepared with a dry KimWipeTM square in each well. These wells were then wetted with water to provide humidity so the buds wouldn’t dry out 4 during the run. Next, each set of 10x or 5x buds, depending on the specific run, were loaded into the proper well and capped with a foam square to seal them in. The 9th well of each tray was not used, to provide a negative freezer control. Every freezer run consisted of a total of 4 trays, each with 8 wells filled with grapevine buds. The trays were loaded into the freezer and plugged into their proper numbered wiring harness which was connected to the voltage data collection system. Then, the freezers were sealed and run according to the pre-programmed DTA settings as described in (Londo & Kovaleski, 2017). 2.3 Peak Calling The LTE peak results were analyzed using two programs, BudProcessor and BudLTE. BudProcessor provided a visual graph which displayed voltage readings of the thermoelectric modules, and is where the LTE peaks at bud time-of-death were called. These peaks varied in size and shape depending on factors such as variety, bud size, and the presence of secondary or tertiary buds. All peaks were evident as a rapid spike in voltage followed by a return to the pre- peak value. Calling of the peak recorded the specific temperature at which the peak was present. Once all peaks were called and committed to the internal database, BudLTE summarized this information and exported the results as a CSV file for further analysis. 2.4 iButton Temperature Logging Twenty-two iButton thermochrons were set up at various points within each vineyard field to record temperature data throughout the growing season. They were set to record the current temperature hourly and were placed at mid-height along the vine trellis. The attempted placement was such that the metal piece was kept out of direct sunlight during the peak sun exposure in the afternoon, to reduce the thermal skew of direct solar radiation. 5 2.5 Statistical analysis All statistical analysis took place using RStudio (v1.3.1073) and the finalized code is available upon request. All graphical figures were produced within the RStudio code. 3. Results 3.1 LTE Measurements The mean LTEs of all clones, separated by variety, can be seen on Figure 2 below. This graph also contains daily weather data sourced from NEWA, which shows the average daily temperature across the collection season. As can be seen, the daily temperature never goes below the LTE point of any given clone, which indicates that all clones were cold hardy enough to survive in this region during the 2020-21 winter season. It is interesting to note that the LTE of all clones shift in relative unison with the average temperatures, following the trend throughout the season until bud break occurs. This is likely due to the mechanics of dormancy, in which the vines adapt to the winter conditions by entering endodormancy, which is a stage at which the buds are responding to a specific environmental signal (chilling) and will no longer grow or develop even if this signal is removed. All vines have a minimum required amount of chilling time, referred to as chill hours (Hentschel, 2020). After this threshold is reached the vines transition to the next stage, ecodormancy. During this stage, the vines stay dormant to avoid damage due to the unfavorable environmental conditions, but can resume growth once the signal is removed (Rubio et al., 2016). As the average temperatures slowly rise starting in mid-February onward, the vines are no longer subjected to consistently low temperatures and begin to lose their ability to withstand low extremes as they transition further through ecodormancy and eventually break bud (Londo et al., 2018). This is represented visually on the graphs as the increase in LTE temperature beginning in mid-February that continues upward throughout the rest of the winter. 6 Figure 2. Graphs showing mean LTE values and standard error (shading) of sampled clones for winter 2020-21, separated by variety. Jagged line is average daily temperature data for this time period sourced from NEWA. Also of note is the shaded band surrounding each clone LTE curve. This shading represents a 95% confidence interval, meaning that if a sample was rerun at the given timepoint, it would be expected to fall within the shaded bounds 95% of the time. This does not apply to samples run during different seasons, as there is variation of daily temperatures and dormancy start which influence the LTE value. However, all clones of each variety fall within the confidence bands of each other for all or most of the season. This indicates that there is not a significant difference 7 between the cold hardiness of clones within any variety, and Cabernet Franc p-value 312-214 1.00000 rather that the variation resides between varieties. Table 3 shows 327-214 0.98270 332-214 0.92326 the relative p-values of a two-way ANOVA and Tukey Honest 3880-214 0.95860 327-312 0.96865 Significant Difference test based on the mean of initial LTE 332-312 0.89134 3880-312 0.93930 values from all collection points. According to the results of this 332-327 0.99352 3880-327 0.99854 statistical test, only one pair of clones, Pinot Noir clones 777 & 3880-332 0.99999 Chardonnay 667, showed significant difference between their initial LTE 95-96 0.84657 Pinot Noir means. When a similar statistical test is run at the variety level 348-115 0.52782 for each collection timepoint, there is significant difference 667-115 0.28066 777-115 0.75381 between all varieties for at least one timepoint throughout the 667-348 0.97878 777-348 0.10760 season (Table 4). This is likely due to varietal differences in 777-667 0.03654 Riesling overall cold hardiness, as well as differences in the rates of 198-110 1.00000 239-110 0.19292 deacclimation in mid and late winter (Londo & Kovaleski, 356-110 0.99987 90-110 0.89490 2017). German-110 0.97231 239-198 0.34007 356-198 0.99995 p-values 90-198 0.94768 Variety Comparisons C1 C2 C3 C4 German-198 0.98077 Chardonnay-Cabernet Franc 0.29231 7.20E-06 0.88819 0.01155 356-239 0.44263 Pinot Noir-Cabernet Franc 0.42934 0.34712 0.98107 <<<0 90-239 0.72212 Riesling-Cabernet Franc 9.17E-05 0.02064 5.40E-05 <<<0 German-239 0.96560 Pinot Noir-Chardonnay 0.99929 0.13036 0.99558 4.19E-05 90-356 0.98402 Riesling-Chardonnay 0.21038 0.04479 1.14E-03 1.14E-02 German-356 0.99367 Riesling-Pinot Noir 0.51278 0.99213 0.00926 0.06800 German-90 1.00000 Table 4. Relative p-values from two-way ANOVA-Tukey HSD test using Table 3. Relative p-values from two-way (LTE~Variety). The pairwise significance of mean initial LTE values at ANOVA-Tukey HSD test using each collection timepoint (C1-C4) are compared. Significant values are (LTE~Clone). The pairwise significance highlighted in red. of mean initial LTE values are compared. 8 3.2 DeAcclimation Rates The deacclimation rates of the vines are governed by their state within the transition from endo to ecodormancy, as well as the pace of their ability to come out of dormancy and break bud at a given timepoint along this transition. This is measured in oC/day, and essentially is a measurement of how quickly the bud is losing cold hardiness as it rises out of dormancy. The rates were calculated by deriving the slopes of each set of DTA runs for a collection. These rates are then displayed in Figure 3 below. As can be seen, the rates vary throughout the winter season, and tend to become faster as spring approaches. This is in agreement with previous studies of deacclimation rates in grapevine and other horticultural crops, as the vines are more endodormant during mid-winter (Dec/Jan) and have transitioned to be mostly ecodormant during late-winter and early-spring (Feb/Mar) (Londo & Kovaleski 2019). At these later timepoints, the vines are simply remaining dormant due to the freezing temperatures, but are able to resume growth relatively rapidly once the freezing temperature signals are removed. As the buds are no longer exposed to freezing temperatures, their level of cold hardiness decreases (i.e. a higher LTE value) at the rate depending on the start date of thawing. Shown by the graphs, Chardonnay and Cabernet Franc tend to have very rapid rates towards the end of the winter season, which makes them very susceptible to spring frost damage, where newly emerged buds are killed by a late frost after they have already come out of dormancy. Riesling tends to have a slower, steadier rate of deacclimation, which lends itself well to the Finger Lakes region where spring frosts can be common. The rate of Pinot Noir is similar to Riesling, however Pinot Noir typically does not have as low of LTE values at a given timepoint, and so is not always able to withstand low temperature extremes during cold winter seasons in this region. (mid-Feb: mean LTE Riesling = -23.73oC / mean LTE Pinot Noir = -22.31oC) 9 Figure 3. DeAcclimation rates of each clone, separated by variety. The points indicate rates measured for each collection, with a line extrapolating the rate between them. 10 3.3 Temperature Logging Data The results of the iButton temperature loggers are shown in Figure 4, as a deviation of hourly recorded data compared to NEWA reported values. The temperatures recorded were either above, at, or below the NEWA value, and the plotted line shows the level of variation (NEWA as y=0). As can be seen, there is quite high deviation in each vineyard from the NEWA values, as well as between vineyard sites. This amount of environmental variation contributes to the variation in LTE values and DeAcc rates for each sampled clone. As they are exposed to Figure 4. Deviations of iButton recorded temperatures from NEWA reported hourly temperature values, separated by vineyard. NEWA temperature is represented by the line y=0. The colored lines show the amount of deviation above or below this temperature. 11 different amounts of ideal chill hours at a given timepoint, each clone, or even reps of the same clone in different vineyards may be at varying stages of endo to ecodormancy transition when sampled at a collection. Due to this, it is extremely difficult to isolate clonal variation as a singular factor when evaluating the overall levels of cold hardiness between clones of a variety. 4. Discussion 4.1 Limitations There were several limitations in this study. Foremost, the lack of each clone in every vineyard site sampled did not allow for a balanced experimental design. This reduced the statistical power of the results and made it more difficult to separate the influence of environment on the observed variation between clones. As each vineyard site had subtle differences in temperature due to factors such as slope, elevation, and distance from the lake, the effect of these environmental variations on the cold hardiness readings of the clones became difficult to distinguish. Additionally, the iButton temperature loggers will need an improved enclosure to reduce solar radiation, as the peak values >10oC on Figure 4 are likely from direct sun exposure in mid- afternoon. This sun exposure may have caused an incorrect reading at a value higher than the realistic temperature. Secondly, time constraints made it challenging to complete regularly sampling throughout the winter. Ideally, the clones would have been sampled weekly to allow for a precise deacclimation rate chart with numerous collection timepoints. A limitation of DTA is that it becomes unreliable when the initial LTE approaches -5oC. This is because the rate of deacclimation occurs too rapidly to measure via LTE freezer runs (13 hour runtime/sample group). As my fourth collection timepoint was taken near the end of the winter season, the starting LTE of most clones 12 was ~-10oC. This caused a discrepancy in the deacclimation rate calculations and showed the rates declining at this timepoint. In reality, the rates may taper off as the vines come out of dormancy, but they should not drop below the previous timepoint as this would indicate a return to endodormancy. This being the case, the fourth timepoint was removed from the calculations and is not present on the DeAcc rate graphs above. Thirdly, a second season of repeat sampling would allow for measures of standard deviation and the formation of error bars on the DeAcc rate graphs. This would provide a more powerful means of statistically comparing the deacclimation rate of the various clones. With only one set of rates, it is challenging to make a comparison, as the rates can and do change in response to the average daily temperatures, as well as other environmental factors. In theory, the rates should only change within a certain set of bounds, determined by the genetic predisposition of each clone. It is with these bounds, represented as error bars, that direct comparisons between deacclimation rates of different clones becomes possible. 4.2 Future Research The results of this study warrant future research into the variation between clones of vinifera grapevines. To more effectively study cold hardiness of the selected clones, a multi-replicate trial should be set up where vineyards with minimal environmental variation (i.e. flat land with no thermal gradients) are planted out with the clonal material and sampled regularly throughout the winter season. Each vineyard should contain all the clones to be tested, and ideally the vineyard sites would be located in multiple different geographical environments. This would allow for both the theoretical elimination of environment as a factor within each vineyard site, as well as the isolation of environment as a contributing variable between vineyard sites. 13 The genetics underpinning clones of grapevine varieties is also an interesting area of potential research. It is possible that there is a range of genetic diversity in and between clones, with some being more closely related within varieties, while others vary greatly due to a more ancient divergence. This spread of diversity could explain the variation in traits such as cluster size/density, flavor profiles, and yield that are commonly seen when growing the clones in situ. Using genetic sequencing and previously developed genetic loci maps for these traits would allow for an in-depth look at the potential variations that exist between clones of each variety. Leveraging commonly known phenotypic differences between two clones, such as a highly visible difference in berry size, may also provide a means of identifying previously unknown loci influencing these traits. 14 REFERENCES Camargo Alvarez, H., Salazar-Gutiérrez, M., Zapata, D. et al. Time-to-event analysis to evaluate dormancy status of single-bud cuttings: an example for grapevines. Plant Methods 14, 94 (2018). https://doi.org/10.1186/s13007-018-0361-0 Hentschel. (2020, March 9). Chilling Hours Help Break Spring Dormancy. University of Illinois Extension. https://extension.illinois.edu/blogs/over-garden-fence/2020-03-09- chilling-hours-help-break-spring-dormancy Londo, J.P., Kovaleski, A.P. Characterization of wild north american grapevine cold hardiness using differential thermal analysis. American Journal of Enology and Viticulture. ajev.2016.16090 (2017). DOI: 10.5344/ajev.2016.16090. Londo, J.P., Kovaleski, A.P. & Lillis, J.A. Divergence in the transcriptional landscape between low temperature and freeze shock in cultivated grapevine (Vitis vinifera). Hortic Res 5, 10 (2018). https://doi.org/10.1038/s41438-018-0020-7 Londo, Jason & Kovaleski, Alisson. Deconstructing cold hardiness: variation in supercooling ability and chilling requirements in the wild grapevine Vitis riparia: Cold hardiness in Vitis riparia. Australian Journal of Grape and Wine Research. 25 (2019). 10.1111/ajgw.12389. Mills, Lynn & Ferguson, John & Keller, Markus. (2006). Cold-Hardiness Evaluation of Grapevine Buds and Cane Tissues. American Journal of Enology and Viticulture. 57. Rubio, S., Dantas, D., Bressan-Smith, R. et al. Relationship Between Endodormancy and Cold Hardiness in Grapevine Buds. J Plant Growth Regul 35, 266–275 (2016). https://doi.org/10.1007/s00344-015-9531-8 15